Model Interpretability refers to how easily machine learning algorithms can understand the processes humans use to reach results. Until recently, artificial intelligence(AI) algorithms were notorious for being a "black box" as it did not offer explanation for its internal processes to help understand how the outcome was produced for regulators and stakeholders.
Some models, such as logistic regression, are considered very simple and highly interpretable, but it becomes increasingly difficult to interpret by adding functionality or using more complex machine learning models such as deep learning. The more interpretable a machine learning model is, the easier it is for someone to understand how a particular decision or prediction has been made.
The features which affect the target variable in order for a classification model.
[Click 'Feature Importance' to see the values selected by the AI model as an important indicator, and additional information are available through other tabs.]
The graph above shows that the churn probability had the greatest impact on the number of products, followed by age, active member, and balance.
[You can check the analysis results, such as the change in target values according to the predicted values.]
[These columns are visualized and provided in a form of various charts such as histogram, distribution chart, and more.]